Toward an Integrative Approach for Making Sense Distinctions.

Word senses are the fundamental unit of description in lexicography, yet it is rarely the case that different dictionaries reach any agreement on the number and definition of senses in a language. With the recent rise in natural language processing and other computational approaches there is an increasing demand for quantitatively validated sense catalogues of words, yet no consensus methodology exists. In this paper, we look at four main approaches to making sense distinctions: formal, cognitive, distributional, and intercultural and examine the strengths and weaknesses of each approach. We then consider how these may be combined into a single sound methodology. We illustrate this by examining two English words, "wing" and "fish," using existing resources for each of these four approaches and illustrate the weaknesses of each. We then look at the impact of such an integrated method and provide some future perspectives on the research that is necessary to reach a principled method for making sense distinctions.

Ontology modularization to improve semantic medical image annotation.

Searching for medical images and patient reports is a significant challenge in a clinical setting. The contents of such documents are often not described in sufficient detail thus making it difficult to utilize the inherent wealth of information contained within them. Semantic image annotation addresses this problem by describing the contents of images and reports using medical ontologies. Medical images and patient reports are then linked to each other through common annotations. Subsequently, search algorithms can more effectively find related sets of documents on the basis of these semantic descriptions. A prerequisite to realizing such a semantic search engine is that the data contained within should have been previously annotated with concepts from medical ontologies. One major challenge in this regard is the size and complexity of medical ontologies as annotation sources. Manual annotation is particularly time consuming labor intensive in a clinical environment. In this article we propose an approach to reducing the size of clinical ontologies for more efficient manual image and text annotation. More precisely, our goal is to identify smaller fragments of a large anatomy ontology that are relevant for annotating medical images from patients suffering from lymphoma. Our work is in the area of ontology modularization, which is a recent and active field of research. We describe our approach, methods and data set in detail and we discuss our results.

Ontology based clinical query extraction.

Knowledge about human anatomy, radiology and diseases that is essential for medical images can be acquired from medical ontology terms and relations. These can then be analyzed using domain corpora to observe statistically most relevant term-relation-term patterns. We argue that such patterns are the basis for more complex clinical search queries and describe our approach for deriving them. These patterns can then be used to support the knowledge elicitation process between the domain expert and the knowledge engineer by providing a common vocabulary for the communication.

Semantic annotation for concept-based cross-language medical information retrieval.

We present a framework for concept-based cross-language information retrieval in the medical domain, which is under development in the MUCHMORE project. Our approach is based on using the Unified Medical Language System (UMLS) as the primary source of semantic data. Documents and queries are annotated with multiple layers of linguistic information. Linguistic processing includes part-of-speech tagging, morphological analysis, phrase recognition and the identification of medical terms and semantic relations between them. The paper describes experiments in monolingual and cross-language document retrieval, performed on a corpus of medical abstracts. Results show that linguistic processing, especially lemmatization and compound analysis for German, is a crucial step in achieving a good baseline performance. On the other hand, they show that semantic information, specifically the combined use of concepts and relations, increases the performance in monolingual and cross-language retrieval.